Multi-class: Spectral-Spatial Temporal Pyramid Network and Multi-class Classifier based Cardiovascular Disease Classification

reehana skSiddique Ibrahim S P^*

• VIT-AP University, Amaravati, India

cardiovascular disease (CVD) represents a major global cause of morbidity and mortality, encompassing conditions such as coronary artery disease, heart failure, and stroke. Early diagnosis of CVD is critical to improve treatment outcomes and prevent disease progression. Leveraging both electrocardiogram (ECG) and phonocardiogram (PCG) data, we propose MCC-CVD, a multi-modality deep learning framework for multi-class CVD classification. Our approach integrates a Spectral Spatial Temporal Pyramid Network (SST-PNet), a Weight Correction Module with Attention Mechanism (WCM-AM), and a novel Multi-class EnDe-CNN classifier to extract complementary features and enhance predictive accuracy. A real-world dataset of 920 patients with 13 clinical parameters was used to evaluate the model. After unified data quality enhancement and normalization, MCC-CVD achieved an average accuracy of 92.4%, F1-score of 0.87, precision of 0.89, recall of 0.85, and an AUC of 0.94, outperforming baseline classifiers such as SVM, Random Forest, and Logistic Regression. These results demonstrate the model’s strong discriminative capability and robustness across multiple CVD subtypes.